https://godbolt.org/z/x5E5cGPTb auto l = []<auto> requires true -> void {}; gcc incorrectly rejects this valid lambda, you can see more details in https://stackoverflow.com/questions/66833334/the-validity-of-lambda-expression-with-omitted-parameter-list-in-c23
Confirmed. I could take a look, unless Jakub wants it.
Are you sure it is incorrectly rejected? http://eel.is/c++draft/expr.prim.lambda.general says: lambda-declarator: lambda-specifiers ( parameter-declaration-clause ) lambda-specifiers requires-clause[opt] So in my reading, if requires-clause is present, the ()s are not optional. Otherwise requires-clause[opt] would need to be in the lambda-specifiers non-terminal or present also after the first lambda-specifiers.
Ah, but in lambda-expression: non-terminal there is another requires-clause[opt] after the < template-parameter-list >. So we need to handle auto l = []<class T> requires true (T t, int n) { }; But during parsing we have parsing of requires clause at that spot after >, and when ()s are omitted, there is code to handle -> at that spot.
This ICE may be caused by not handle this form. template <typename...> concept C = true; auto l = []<typename... Ts> requires (C<Ts> && ...) -> void {}; https://godbolt.org/z/vo8xPd4hY <source>:3:48: internal compiler error: Segmentation fault 3 | auto l = []<typename... Ts> requires (C<Ts> && ...) -> void {}; | ^~~ 0x1cfca39 internal_error(char const*, ...) ???:0 0x940286 convert_generic_types_to_packs(tree_node*, int, int) ???:0 0x8e126d c_parse_file() ???:0 0xa60292 c_common_parse_file() ???:0 Please submit a full bug report, with preprocessed source if appropriate.
Created attachment 50592 [details] gcc11-pr99850.patch Untested fix.
The master branch has been updated by Jakub Jelinek <jakub@gcc.gnu.org>: https://gcc.gnu.org/g:784de5292c34e287c848b382b431599b818ea76e commit r11-8210-g784de5292c34e287c848b382b431599b818ea76e Author: Jakub Jelinek <jakub@redhat.com> Date: Fri Apr 16 09:34:26 2021 +0200 c++: Fix up C++23 [] <...> requires primary -> type {} parsing [PR99850] The requires clause parsing has code to suggest users wrapping non-primary expressions in (), so if it e.g. parses a primary expression and sees it is followed by ++, --, ., ( or -> among other things it will try to reparse it as assignment expression or what and if that works suggests wrapping it inside of parens. When it is requires-clause that is after <typename T> etc. it already has an exception from that as ( can occur in valid C++20 expression there - starting the parameters of the lambda. In C++23 another case can occur, as the parameters with the ()s can be omitted, requires C can be followed immediately by -> which starts a trailing return type. Even in that case, we don't want to parse that as C->... 2021-04-16 Jakub Jelinek <jakub@redhat.com> PR c++/99850 * parser.c (cp_parser_constraint_requires_parens) <case CPP_DEREF>: If lambda_p, return pce_ok instead of pce_maybe_postfix. * g++.dg/cpp23/lambda-specifiers2.C: New test.
Fixed.